How Alterra Mountain Company Performed a Total Data Center Refresh with Low Organizational Impact
Based out of beautiful Denver, Colorado, Alterra Mountain Company is an organization founded on a single premise: We want to encourage and empower your passion for outdoor adventure.
The company began by joining together a number of popular destination resort companies under one roof. Our goal is to continue honoring each destination's unique character and authenticity, plus create a consistent and enjoyable experience that enables our visitors to take legendary adventures at any time.
Our oldest property is the Winter Park Resort, which opened its doors back in 1939 and also holds the distinction of being the first ski resort in Colorado. We operate resorts in Washington state, Utah, West Virginia, Vermont, California, and parts of Canada. The property in Ontario, Canada—Blue Mountain—is the third busiest ski resort in the entire country.
Soon after I joined one of the companies that became Alterra, one of our top priorities was to refresh our aging equipment. Since we couldn’t just transition our applications to the cloud in a cost-effective manner, the decision was made to replace our existing on-premises environment entirely. We started with our secondary data center using equipment that would also be workable in our resort sites. We later focused on growing that environment to handle migrating workloads from the primary data center which was targeted for decommission.
What Does the Future Look Like?
Coming from a managed services background, I recognized immediately that our hardware requirements were rather modest. The reference architecture that our data center and resorts had previously undertaken was one of the typical triple-tier hardware architecture: compute, storage, and networking. Because WAN reliability left something to be desired at the time, the overall goal of the time was to make the hardware bulletproof using top-tier vendors with a best of breed solution.
This added complexity that, in present day, was no longer required. Hyperconverged infrastructure solutions eliminate that complexity. With our modest compute and storage requirements, however, I couldn't justify the complexity of the old architecture and the requisite SAN environment. It would be too taxing on a resort staff that already had a plethora of other responsibilities.
The decision to stay on-premises was further solidified when we considered our industry. In the hospitality industry, we have some elderly software that’s not necessarily designed to scale horizontally. When you also need to run 24/7/365, and realize you're paying by the hour, cloud computing will be more expensive than on-premises hardware.
That, coupled with the budget pressure we were already under, made this decision fairly straightforward: We needed to pursue a hyperconvergence infrastructure.
In the early stages, we considered Nutanix. They seemed to be the market leader in the space, but I felt like I had to do my due diligence and look at a few competitors. Quickly, I realized that the entry footprint for HPE SimpliVity (at the time, just SimpliVity) was better suited for our needs. Unlike Nutanix, HPE SimpliVity didn't require the rip-and-replace approach: We could augment our existing compute as needed. At the same time, some of our resorts didn't require the larger compute footprints that Nutanix would demand. In some situations, I didn't envision the need for five (or four) nodes at a single site, which was the minimum footprint we would be dealing with had we gone with Nutanix or some of the others.
Nutanix originally had the lead because that was the solution everyone was talking about. But because of the flexibility and lower barrier to entry, HPE SimpliVity was the right option for us.
A Rapid Expansion
Our project began with two medium-sized SimpliVity nodes that replaced our existing, aging infrastructure at the smaller of our two data centers. Through this, we were able to reduce roughly half of a cabinet worth of gear with 4U of the two nodes. As we continued with the process of decommissioning our larger gear, we saw that our environment was growing. During our discussions, we realized that, based on our needs, we should have started with larger nodes.
Lesson learned, I suppose. Had we known then what we know now, we would have started with two large nodes so that, as we deployed more nodes to handle larger workloads, we could grow at the same increment. (One of the enhancements that HPE has introduced is the ability to mix nodes of different sizes—an option we didn’t have in the days before HPE bought SimpliVity).
This all began in April of 2016 with just those two nodes. A year later, we'd grown to 11 nodes. The environment at the smaller data center was growing, which allowed us to get rid of roughly a dozen full cabinets at our larger data center. This was a huge cost savings for us, as we were paying over a thousand dollars a cabinet. That alone was a massive justification that we were on the right track.
At this point, we were certain this technology would work. Based on our early success, I was able to convince my leadership and colleagues to ramp up our expansion into other resorts.
Making the Case for Continued Success
In August of 2017, we deployed three HPE SimpliVity nodes with all-flash, solid state drives at one of our eastern resorts just prior to the beginning of ski season. The pitch was a simple one: Because this solution was working in our data centers, it would be perfect for some of our secondary sites, too. The gear could go back to being locally deployed at the edge, but the local resort staff wouldn't be tasked with taking care of it.
We chose our Winter Park, Colorado resort as the first site to add to the Denver federation, as geographically, they were the closest to our main office. In addition, they didn't need any new infrastructure to participate, whereas all of the other resorts would. We could just plug in and go.
To help this go smoothly, we offered training where people could sign up for courses or participate in Webex sessions so I could convey all critical information. I condensed two days' worth of training into four hours' worth of presentations. At the end, the resort staff members who participated felt I had covered everything they needed.
But as we planned this multi resort expansion, we ran into some hurdles. Due to some of organizational changes, this resort upgrade plan had to be re-sold to layers of management which were unfamiliar with the plan and technology. In addition, those engineers at the resorts had to buy into the strategy and sizing. But we were always able to sell our vision.
The Start of a Streamlined Organization
Our long-term goal for these initiatives involves significantly streamlining our organization by further extending our new hyperconverged federation. Right now, things are still foundational—we can't check that box quite yet. But even still, our IT performance today is exceptionally better than it was when our project began.
We now have 40+ nodes live at various locations, and I can manage everything from a single vCenter dashboard. But more importantly, that foundation is in place so that engineers at one resort can easily see what people at another resort are doing. They can start creating common practices and configurations.
The data efficiency gains we've seen with HPE SimpliVity and inline dedupe and compression for capacity savings have been enormous. Big Bear is one of our smaller sites, for example, and that one's achieving a 237:1 data efficiency rate. Other sites are at 166:1 or even 195:1. It depends on the site and where they are in the project, but the results speak for themselves.
There's still a certain amount of ownership in place at each environment, since all the resorts want to maintain control over their local resources. But if there is ever a situation where one resort needs help from another, they can easily collaborate. As management decides on a future organizational structure, the infrastructure is in place to handle changes of roles and responsibilities.
The feedback we've received throughout the organization has been overwhelmingly supportive. In Winter Park, for instance, I had someone say to me: "We put it in, and we haven't touched it since we set it up." As an architect, that’s incredibly reassuring to hear.
Another reassuring moment happened recently when we ran into a problem at one location with an application and the vendor recommended our team reinstall the software. But our guys didn’t want to do that—they’d rather just restore from a backup. A virtual machine restore process that would have taken three hours before took ten seconds and a few simple clicks with HPE SimpliVity. Having the app down for three hours would have been a big deal. Thankfully, with HPE SimpliVity, that’s a thing of the past.
When you consider the tremendous amount of expansion we've had over this short period of time, it's obvious that HPE SimpliVity was the best solution. HPE SimpliVity has allowed us to grow without needing to invest in a huge amount of capital at each location. And using VMware vCenter meant I wouldn't have to train and retrain everyone at each site, which would have required a minimum of four days, per location.
Using HPE SimpliVity has also allowed us to pull off this massive refresh, all with low impact on the business and a modest time investment. We wouldn’t have accomplished this with another solution.
Our goal in forming Alterra was to join all of these companies under the same roof. We wanted to unify to bring you the best outdoor experiences. Our progress on the IT side means we’re better equipped than ever to take your mind off of your troubles and onto the slopes.